Application of Structure-Based Design and Parallel Chemistry to Identify a Potent, Selective, and Brain Penetrant Phosphodiesterase 2A Inhibitor

Article abstract of DOI:10.1021/acs.jmedchem.7b00397

Phosphodiesterase 2A (PDE2A) inhibitors have been reported to demonstrate in vivo activity in preclinical models of cognition. To more fully explore the biology of PDE2A inhibition, we sought to identify potent PDE2A inhibitors with improved brain penetration as compared to current literature compounds. Applying estimated human dose calculations while simultaneously leveraging synthetically enabled chemistry and structure-based drug design has resulted in a highly potent, selective, brain penetrant compound 71 (PF-05085727) that effects in vivo biochemical changes commensurate with PDE2A inhibition along with behavioral and electrophysiological reversal of the effects of NMDA antagonists in rodents. This data supports the ability of PDE2A inhibitors to potentiate NMDA signaling and their further development for clinical cognition indications.

Full text of DOI:10.1021/acs.jmedchem.7b00397

Article
Application of Structure-Based Design and Parallel Chemistry to Identify
a Potent, Selective, and Brain Penetrant Phosphodiesterase 2A Inhibitor
Christopher J. Helal, Eric P Arnold, Tracey L Boyden, Cheng Chang, Thomas A. Chappie, Kimberly F.
Fennell, Michael D Forman, Mihaly Hajós, John F. Harms, William E. Hoffmann, John M. Humphrey,
Zhijun Kang, Robin J. Kleiman, Bethany L. Kormos, Che-Wah Lee, Jiemin Lu, Noha Maklad,
Laura McDowell, Scot Mente, Rebecca E O'Connor, Jayvardhan Pandit, Mary Piotrowski, Anne
W. Schmidt, Christopher J. Schmidt, Hirokazu Ueno, Patrick R. Verhoest, and Edward X. Yang
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 02 Jun 2017
Downloaded from http://pubs.acs.org on June 2, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Application of StructureꢀBased Design and Parallel
Chemistry to Identify a Potent, Selective, and Brain
Penetrant Phosphodiesterase 2A Inhibitor
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
a
a
d
Christopher J. Helal,* Eric P. Arnold, Tracey L. Boyden, Cheng Chang, Thomas A.
b
a
a
b
b
Chappie, Kimberly F. Fennell, Michael D. Forman, Mihaly Hajos, John F. Harms, William
b
a
a
b
b
E. Hoffman, John M. Humphrey, Zhijun Kang, Robin J. Kleiman, Bethany L. Kormos, Cheꢀ
a
a
a
a
b
a
Wah Lee, , Jiemin Lu, Noha Maklad, Laura McDowell, Scot Mente, Rebecca E. O’Connor,
a
a
a
b
Jayvardhan Pandit, Mary Piotrowski, Anne W. Schmidt, Christopher J. Schmidt, Hirokazu
a
b
a
Ueno, Patrick R. Verhoest, and Edward X. Yang
a
Pfizer Worldwide Research and Development, Eastern Point Road, Groton, CT 06340
Pfizer Worldwide Research and Development, 610 Main Street, Cambridge, MA 02139
b
KEYWORDS: phosphodiesterase, parallel chemistry, structureꢀbased drug design, protein xꢀray
crystallography, brain penetration, schizophrenia, cognition, radio ligand, cyclic guanosine
monophosphate, ketamine, MKꢀ801, radial arm maze, electroencephalogram
ABSTRACT: Phosphodiesterase 2A (PDE2A) inhibitors have been reported to demonstrate in
vivo activity in preclinical models of cognition. In order to more fully explore the biology of
PDE2A inhibition we sought to identify potent PDE2A inhibitors with improved brain
penetration as compared to current literature compounds. Applying estimated human dose
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
calculations while simultaneously leveraging synthetically enabled chemistry and structureꢀbased
drug design has resulted in a highly potent, selective, brain penetrant compound 71 (PFꢀ
05085727) that effects in vivo biochemical changes commensurate with PDE2A inhibition along
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
with behavioral and electrophysiological reversal of the effects of NMDA antagonists in rodents.
This data supports the ability of PDE2A inhibitors to potentiate NMDA signaling and their
further development for clinical cognition indications.
Introduction
Phosphodiesterases (PDEs) are enzymes that catalyze the hydrolysis of the key second
messenger cyclic nucleotides cAMP and cGMP to their respective acyclic variants (AMP, GMP).
PDE inhibitors have demonstrated therapeutic potential for a range of conditions such as
pulmonary hypertension (sildenafil, PDE5 inhibitor), erectile dysfunction (sildenafil, vardenafil,
tadalafil, PDE5 inhibitors), chronic obstructive pulmonary disease (roflumilast, PDE4 inhibitor),
psoriasis and psoriatic arthritis (apremilast, PDE4 inhibitor), and heart failure (milrinone, PDE3
inhibitor) reflecting the wide range of biological functions regulated by cyclic nucleotide
signaling. The abundant expression of PDEs in the human brain, coupled with a requisite role
for cyclic nucleotides in neurotransmission and synaptic plasticity, has prompted efforts to
discover PDE selective inhibitors with brain penetration suitable for treating neurological and
psychiatric disorders. Brain penetrant clinical candidates have been disclosed as inhibitors for
1
,2
PDE10A for the treatment of schizophrenia and Huntington’s Disease as well as PDE9A
3,4
inhibitors for treatment of cognitive disorders.
Inhibitors of PDE2A have been reported to demonstrate activity in a variety of preꢀclinical
models that are suggestive of cognitive improvement. This includes enhanced performance in
ACS Paragon Plus Environment

Page 3 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
rodent novel object recognition and social recognition tasks and a reversal of performance
deficits produced by treatment with the NꢀmethylꢀDꢀaspartate (NMDA) antagonist MKꢀ801 in
5
the alternating Tꢀmaze. These behavioral effects are consistent with reports of enhanced
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
hippocampal longꢀterm potentiation, a measure of synaptic plasticity thought to underpin some
forms of learning and memory. Additional data suggests that PDE2A inhibitors can reverse
6
anxiolytic behavior in mice via cGMPꢀmediated processes.
This aggregate information
encouraged our investment in the discovery of potent, selective, and brain penetrant PDE2A
inhibitors to more effectively probe the potential of this class of molecules to treat a range of
neurological disorders. For medicinal chemistry lead optimization, tracking binding efficiency
and minimizing estimated human dose were supported by structureꢀbased drug design and
parallel synthesis strategies.
Results
The potent imidazotriazinone PDE2A inhibitor 1 (BAY 60ꢀ7550), PDE2A IC = 0.5 nM (our
50
5
assay), 4.7 nM (literature assay) ) has most often been used to explore the pharmacology of
7
8
PDE2A inhibition (Table 1). A ligand efficiency (LE) of 0.37 (based upon our assay) indicated
9
a high binding efficiency, and a lipophilic ligand efficiency (LLE) of 6.7 placed 1 in a range
comparable to clinical candidates with a desirable balance between potency and lipophilicity.
Selectivity of 1 for PDE2A as compared to PDE1B (PDE1B IC /PDE2A IC ) was ca. 500ꢀfold
5
0
50
and significantly higher for all other PDEs (PDE3A/4D/5ꢀ11). ADME measurements revealed
to have high human liver microsome (HLM) clearance as represented by the apparent intrinsic
clearance (HLM CL ), short halfꢀlife (t ) in rat liver microsomes (3 min) and significant efflux
1
int
1/2
in a Pꢀglycoprotein (PꢀgP) multidrug resistance protein (MDR1) over expressing cell line which
10
is indicative of high probability of impaired brain penetration.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1: Potency, selectivity, and in vitro ADME properties of PDE2A inhibitor 1 (1)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
PDE2A IC (nM)
0.5
5
0
534x (PDE1B)
>
1000x
(PDE3A/4D/5
11)
PDE Selectivity
ꢀ
LE (LipE)
0.37 (6.9)
219
a
HLM, CLint
b
MDR1 BA/AB
16.2
#
RB
10
a) values represent the HLM Clint,apparent scaled (ꢀΛ/min/kg).; b) AB = Apical to basolateral; BA =
Basolateral to apical; c) cpKa = calculated acid dissociation constant;
While 1 is a sufficient PDE2A inhibitor tool compound for in vitro studies, in vivo exposure both
peripherally and in brain would be predicted to be poor. Indeed, low in vivo brain exposure has
1
1
been reported recently and was corroborated in similar experiments by us (see Supporting
Information).
An xꢀray crystal structure of 1 bound to PDE2A was determined which showed the binding
mode in the active site, and allowed development of a plausible hypothesis to explain the
selectivity profile of 1 (Figure 1). This crystal structure, determined independently, is
12
isomorphous to the one described by Zhu et al, with an r.m.s. deviation of 0.39 angstroms for
ACS Paragon Plus Environment

Page 5 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
all backbone Cα atoms. We highlight here the critical features of the structure that we exploited
in our subsequent medicinal chemistry design strategy. An induced pocket is seen, formed by the
movement of Leu770, which accommodates binding of the propylphenyl group. The induced
pocket, lined by mainly hydrophobic residues (Leu770, Leu809, Ile870, His773 and Leu774) had
not previously been observed in any PDE structure, and is absent in all the previous PDE2
crystal structures we had determined. We hypothesized that compounds capable of inducing this
conformational change in PDE2 would retain the relatively high PDE selectivity of 1. The
conserved glutamine at the back of the binding pocket (Gln859) makes a pair of hydrogen bonds
with the imidazotriazinone core C=O(NH). Gln859 assumes the same rotamer as it would if
bound to cGMP. A water molecule that is bound by the sidechain of Tyr655 and backbone
carbonyl of Asp808 donates a hydrogen bond to Nꢀ6 of 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Xꢀray structure of 1 bound to PDE2A cyclic nucleotide binding site (hydrogen bond
distances in angstroms). Polar hydrogen atoms used in in silico docking calculations have been
indicated in their modeled positions in this and all other figures.
Based upon the challenges with modifying 1 to attain the desired alignment of potency,
selectivity, and ADME parameters, alternative PDE2A inhibitor lead matter was sought. Highꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
throughput screening identified two novel series of PDE2A inhibitors: a pyrazolo[3,4ꢀ
d]pyrimidine series (2) and a [1,2,4]triazolo[4,3ꢀa]pyrazine series (3) (Tables 2 and 3). Both 2
and 3 lacked the HBDs in 1 that potentially contributed to the observed efflux in the MDR1
BA/AB assay, resulting in poor in vivo brain penetration. Our goal was to achieve equivalent
potency to 1 and equally high (or higher PDE selectivity) with these new series during lead
development while maintaining the CNSꢀpenetration favoring physical properties.
Table 2. Potency, selectivity, in vitro ADME and physical properties of series 2 and 3 identified
via a PDE2A high throughput screen
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
O
N4
N2
N7
N1
2
146.4
0.42 (3.9)
123
3
389.6
0.37 (2.7)
>287
1.2
PDE2A IC (nM)
5
0
LE (LipE)
HLM Clint
MDR1 BA/AB
1.6
#
RB
3
7
estimated human
8787
>44142
a
dose (mg/day)
a) estimated human dose (mg/day) = (10 x PDE2A IC50 x HLM CLint x 24h x 70 kg) / (MDR1 BA/AB); see
Equation 1.
a
Table 3. PDE Selectivity Profiles of 2 and 3
ACS Paragon Plus Environment

Page 7 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
42
7
B
3A
23
6
4D
23
29
5A
3
71
6
17
69
7B
69
16
8B
>615
206
9A
>615
>231
10A
0.05
5
11A
0.3
0.4
2
3
a) PDE selectivity is measured as the ratio of PDEX IC / PDE2A IC .
50
50
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Another goal of the lead optimization process was to minimize the estimated human dose since
1
3
higher human drug dosages are strongly correlated with idiosyncratic adverse drug reactions.
The equation used for estimating the human dose for a brain targeted compound is shown in
Equation 1 (see Supporting Information for derivation).
ꢀꢁꢂꢃꢄ(ꢅꢆ/ꢇꢈꢉ) = 10 ∗ ꢊꢀꢋ2ꢌꢄꢍꢎ50 ∗ ꢏꢈꢐ ∗ ꢑꢒꢓꢄꢎꢒ_ꢔꢕꢖ ∗ ꢗꢘ ∗ ꢓꢀꢙꢄꢗꢌ/ꢌꢗ
Equation 1. Human dose equation for a brain targeting compound based upon in vitro data. Legend: PDE2A
IC : compound in vitro potency at the PDE2A enzyme; Tau: dosing duration (24 h was used assuming once a day
50
dosing); HLM CLint was used as a surrogate for CLint,u (unbound intrinsic clearance); BW: typical human body
weight, 70 kg was used; MDR1 BA/AB: The BA/AB ratio from the MDR1 assay was used as a surrogate for brain
availability.
In the case of 2, the estimated human dose projected by using this method afforded a value of
8787 mg/day. For 3 the estimated human dose was >44142 mg/day. In both cases it was clear
that significant reductions in PDE2A IC and HLM Cl would be necessary while minimizing
5
0
int
MDR1 BA/AB values to get to the desired target of estimated human dose <100 mg/day.
In order to understand the critical binding elements for each series, PDE2A protein xꢀray crystal
structures were sought with compounds 2 and 3 bound. Compound 3 yielded a highꢀresolution
crystal structure with PDE2A (Figure 2a), but attempts at obtaining an xꢀray crystal structure of 2
with PDE2A were not successful. Efforts with a related compound 4 (Table 4), with a similar
PDE inhibition profile as 2 that was identified via initial lead optimization efforts, were
successful (Figure 2b).
Table 4: PDE2A potency and PDE selectivity profile of 4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
PDE2A IC (nM): 66.6
5
0
Fold PDE Selectivity (PDEX
IC / PDE2A IC )
5
0
50
1B
64
3B
4D
112
9A
5A
11
6
4
58
7B
8B
10A 11A
1
04 >1000 >1000 0.1
3
a)
b)
ACS Paragon Plus Environment

Page 9 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
c)
d)
Figure 2. a) Xꢀray crystal structure of 3 bound to PDE2A cyclic nucleotide binding site
hydrogen bond distances in angstroms); b) Xꢀray crystal structure of 4 bound to PDE2A cyclic
(
nucleotide binding site (hydrogen bond distances in angstroms); c) Overlap of PDE2A xꢀray
crystal structures of 3 (cyan) and 1 (blue); d) Overlap of PDE2A xꢀray crystal structures of 4
(
yellow) and 1 (blue).
The core heteroaromatic ring systems of 3 and 4 occupy roughly the same space as the
corresponding imidazotriazinone ring system in 1 (Figures 2c and 2d). For a dualꢀspecificity
phosphodiesterase such as PDE2A, Gln859 is proposed to act as a switch, capable of freely
rotating so as to form hydrogen bonds either with the exocyclic amino group of cAMP or the
14
exocyclic carbonyl oxygen of cGMP. In the structures of 3 and 4, Gln859 must be in a flipped
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
orientation relative to that seen in the structure of 1 to allow for optimal proteinꢀligand
interactions with Nꢀ7 of 3 and Nꢀ5 of 4.
A second major difference is that 3 and 4 do not induce the conformational change in Leu770
which leads to the formation of the selectivity pocket, occupied by the phenpropyl moiety in 1.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In all three crystal structures, substituents on the core extend towards the protein surface. The
methoxy phenyl groups in 3 and 4 extend out of the substrateꢀbinding site, and are oriented
approximately perpendicular to the core ring, making an edgeꢀtoꢀface interaction with Phe862.
These hydrophobic groups occupy a shallow channel on the surface of the protein, sandwiched
between the side chains of Phe862 and Ile866 on one side and Phe830 and Met847 on the other
side (Figures 2a and 2b). An important observation is that the methoxy phenyl substituents on 3
and 4 would clash with Leu770 if it were to assume the ‘out’ rotamer needed to create the
selectivity pocket (Figures 2c and 2d).
Another common interaction is that of a core ring nitrogen (1 – N6, 3 – N7, 4 – N2) with the
structural water molecule which interacts with the sidechain of Tyr655 and the backbone
carbonyl of Asp808. All three compounds also have substituents (1 ꢀ 2ꢀhydroxyethyl, 3 and 4 –
iBu) which occupy a pocket that is filled by the ribose of cyclic nucleotides.
Binding features unique to 1 and 4 relative to 3 are the interaction of C(4)=O and Nꢀ7,
respectively, with Gln812. Comparatively, Nꢀ1 of 3 binds to a water molecule in the back of the
pocket whereas both 1 and 4 displace the water with a methyl group (1 – C5, 4 – N1).
The synthetic route to the pyrazolopyrimidine series (Scheme 2) entailed acylation of malonitrile
(5) with isobutyryl chloride (6) as the first step to afford 7. Oꢀalkylation yielded 8 followed by
condensation with Nꢀmethyl hydrazine to afford aminopyrazole 9. Acidic hydrolysis of the
ACS Paragon Plus Environment

Page 11 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
nitrile to primary amide (10) followed by condensation with formic acid yielded
pyrazolopyrimidinone 11, that was chlorinated with POCl to yield penultimate chloro
3
pyrazolopyrimidine 12. S Ar reaction of 12 with amines afforded analogs of general structure
N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
3.
O
NC
Cl
CN
+
a
b
c
NC
d
NC
NC
H N
5
N
N
2
N
OH
O
(7)
(6)
(5)
(4)
O
CN
CN
N
7
H N
2
H2N
8
9
10
e
(3)
6
R'
R''
O
N
N
N
Cl
N
g
1)
f
HN
N
N
N
N
N
N
(
(2)
N
N
13
1
2
11
Scheme 2. Synthetic route to pyrazolo[3,4ꢀd]pyrimidine targets of general structure 13 with key
diversity points highlighted in green. a) NaH, THF; acid chloride, 76% b) Me SO , NaHCO ,
2
4
3
dioxane, water, 56% c) MeNHNH , EtOH, 35% d) H SO , 68% e) HCO H, 71% f) POCl ,
2
2
4
2
3
crude g) R’R’’NH, K CO , THF, H O (75% for 2).
2
3
2
The synthetic route to the triazolopyrazines started with a selective S Ar reaction between an
N
alkoxide 14 and 2,6ꢀdichloropyrazine (15) in Step 1 to afford the ether 16 (Scheme 3). A second
SNAr with hydrazine yielded 17 (Step 2). A oneꢀpot hydrazone formation with isobutyraldehyde
(18) followed by oxidative triazole formation with phenylꢀiodoacetate in Step 3 yielded 19. Of
particular note in contrasting the synthetic routes to 13 and 19, is that the isobutyl group in 19,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
which overlays with that of 13 in the ribose region and with the potential to access the selectivity
pocket, is introduced in the last step of the synthetic sequence and is amenable to variation via
parallel synthesis which is not the case for the isobutyl vector in 13.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Synthetic route to [1,2,4]triazolo[4,3ꢀa]pyrazine targets of general structure 3 with
key diversity points highlighted in green. a) KOtBu, CH Cl b) NH NH , EtOH c) i. 17 + 18,
2
2
2
2
ClCH CH Cl; ii. (Diacetoxy)iodobenzene.
2
2
The starting hypothesis was that growing the isobutyl group of either 2 or 3 could lead to
formation and filling of the PDE2A induced pocket with significant increases in both PDE2A
potency and overall selectivity over other PDEs while maintaining low MDR1 BA/AB. The
highly enabled nature of the isobutyl in 3 led to a parallel synthesis effort in this series. A key
concern from a structural perspective during the design of the initial parallel synthesis compound
sets was the observed movement of Leu770 towards the 2ꢀmethoxyphenyl group of 3 when the
induced pocket is formed and the potential for negative proteinꢀligand interactions which could
reduce potency as previously mentioned (Figure 2d).
ACS Paragon Plus Environment

Page 13 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In order to address this concern, removal of the ether substituent in the initial parallel synthesis
to afford compounds of general formula 21 was undertaken, realizing that potency losses could
be addressed in followꢀup compounds by reincorporating ether substituents (Scheme 4).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 4. Oneꢀstep synthetic scheme to vary vector headed towards induced pocket. a) i. 20 +
1
8, ClCH CH Cl; ii. (Diacetoxy)iodobenzene
2 2
More than 2000 virtual compounds were enumerated by using the available aldehydes in the
Pfizer monomer file which were docked into the PDE2A xꢀray crystal structure. Compounds
were chosen for synthesis by using the following criteria:
1
. Potential for accessing deep into the induced pocket in a manner analogous to the pendant
phenyl of 1.
2
. Possible hydrogen bond interactions were observed in the ribose region, similar to the
hydroxyl of 1.
If there were highly similar aldehyde monomers available, a representative was chosen to
minimize duplication and maximize diversity of the parallel synthesis library.
Compounds were removed from consideration if clear negative interactions with the
protein existed.
This lead to 484 compounds being chosen for actual synthesis in a parallel format. From
this set, 283 were successfully purified by HPLC (283/484 = 58% success rate) and were
screened in the PDE2A inhibition assay. From this effort one compound (22) was
identified with ꢀM potency and favorable physicochemical properties (Table 5).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 5. PDE2A potency, lipophilicity and binding efficiency of 22
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
22
PDE2A IC50
2107.4
(nM)
clogP
LE
2.2
0.27
2.5
LipE
Computational docking of 22 in the PDE2A protein and overlay with 1 (Figure 3) suggested that
the trifluoromethyl group was positioned in the same space as the pendant phenyl of 1 and that
the pyrazole could engage in hydrogenꢀbonding interactions with tightly bound waters around
the catalytic metals in a similar manner to the hydroxyl of 1. The preferred pyrazole tautomer
was not clearly defined, with the pyrazole having the potential to be either a hydrogen bond
donor or acceptor. In order to further optimize this promising hit, three different approaches
were explored: 1) reintroduction of the ether at the 5ꢀposition of the [1,2,4]triazolo[4,3ꢀ
a]pyrazine core, 2) variation of the pyrazole nitrogen substituents to lock out a specific tautomer,
and 3) evaluation of alternative aryl groups to the paraꢀtrifluoromethyl phenyl.
ACS Paragon Plus Environment

Page 15 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Modeled pose of 22 (cyan) overlayed with 1 (blue).
In the next rounds of parallel synthesis, two different parallelꢀenabled synthetic routes were used.
The 3ꢀstep synthetic route in Scheme 3 was employed to introduce different alcohols (14a-b) at
the 5ꢀposition to see if potency could be increased while concurrently using pyrazoloꢀaldehydes
(23) from the Pfizer monomer store affording compounds of structure 24 (Scheme 5).
R'
Ar
N
N
R
N
R
N
R
N
Cl
O
O
O
a
b
c
N
N
N
N
N
N
(3)
(2)
NH
2
(1)
N
Cl
+
Cl
N
H
15
16
17
24
+
O
Ar
R
OH
H
R =
N R'
23
14a: 2-(2-methoxyphenyl)ethyl
N
14b: n-PrOH
Ar = substituted phenyl
R' = H or Me
Scheme 5. Parallel synthesis route to introduce different ether substituents and substituted
pyrazoles using methods of Scheme 3. a) KOtBu, CH Cl b) NH NH , EtOH c) i. 17 + 18,
2
2
2
2
ClCH CH Cl; ii. (Diacetoxy)iodobenzene.
2
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In the second library, the aryl group attached to the pyrazole was varied via a 4ꢀstep sequence
wherein the ether substituent was fixed as nꢀpropyl (Scheme 6). Starting with iodopyrazole 25,
Suzuki coupling with aryl boronates afforded 26. Selective Cꢀ4 iodination gave 27 that could be
carbonylated in the presence of triethylsilane to yield 23. Condensation of 23 with 17a followed
by oxidative cyclization gave analogs 24a. Compounds derived from both routes were
resynthesized on larger scale for more thorough characterization via the route shown in Scheme
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7.
Scheme 6. a) ArB(OH) , Pd(dppf)Cl , Cs CO , DMF b) NIS, MeCN c) Pd(OAc) , dppf. CO,
2
2
2
3
2
TEA, SiHEt d) 2ꢀhydrazinylꢀ6ꢀpropoxypyrazine 17a, EtOH; (diacetoxyiodo)benzene, EtOH
3
ACS Paragon Plus Environment

Page 17 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
R
Cl
Cl
O
R
R
a
b
N
N
N
N
N
N
NH2
N
N
N
N
N
H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
28
29
19
+
O
H
R
18
Scheme 7. Reꢀsynthesis route for key library compounds. a) i. 28 + 18, ClCH CH Cl; ii.
Diacetoxy)iodobenzene b) ROH, NaH, THF
2
2
(
Two different alcohol starting materials were employed in Scheme 5: 2ꢀ(2ꢀ
methoxyphenyl)ethanol which is in 3, along with the lower molecular weight alcohol nꢀpropanol
which was expected to have reduced steric interactions with Leu770 due to its smaller size. As a
direct comparator to 3, 3ꢀisobutylꢀ5ꢀpropoxy analog 30 was prepared and a 5ꢀfold decrease in
PDE2A potency was observed but LipE remained essentially equivalent due to reduced
lipophilicity (Table 6). LE increased, indicating more optimal interactions of 30 with protein
relative to 3 on a per atom basis.
Table 6. PDE2A potency, physical properties and binding efficiencies of 3 and 30
Cpd. ID
R
PDE2A
clogP
2.7
LE
LipE
2.5
IC (nM)
5
0
3
0
5323.8
0.43
3
389.6
3.7
0.37
2.7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Reincorporating the 5ꢀether substituent made a significant impact on potency (Table 7). For
example, the nꢀpropoxy group (31) afforded a >10ꢀfold increase in potency over the parent 22.
Lipophilicity was significantly higher for 31 vs. 22 (clogP = 3.8 vs. 2.2), yet LipE values were
equivalent and LE was improved for 31. As the ether substituent size and lipophilicity increased
in 32, potency values increased slightly relative to 31 with concomitant decreases in LE and
LipE indicating more efficient binding of 31.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 7. Impact of 5ꢀposition substitution on PDE2A potency and binding efficiency measures
PDE2A IC
5
0
Cpd. ID
R
H
a
clogP
LE
LipE
(nM)
2
3
2
1
2107.4
2.2
3.8
0.27
0.35
2.5
3.1
115.0
3
2
24.4
4.8
0.30
2.8
a) PDE2A IC data is from resynthesized samples, not from initial library compounds.
5
0
The next key finding was that methylation of the pyrazole nitrogen adjacent to the aromatic
ring afforded >10ꢀfold increase in PDE2A inhibitory potency when compared to 33 and 34
(Table 8). In order to provide a second example of this trend, 35, the Nꢀmethylated version of
library compound 31, was prepared and the matched pair (31 and 35) showed the same relative
potency increase. This result was critical in that not only was potency improved, but a HBD was
removed that resulted in reduced MDR1 BA/AB efflux. The increased lipophilicity due to the
methylation, however, increased HLM Clint.
ACS Paragon Plus Environment

Page 19 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 8. Impact of pyrazole Nꢀsubstitution on PDE2A potency, MDR1 BA/AB, and HLM Clint
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
PDE2A
IC ratio
R=H/Me
MDR1
BA/AB
5.6
5
0
Cpd. ID
R
R’
a
HLM Clint
IC (nM)
5
0
3
3
3
4
H
Me
H
Me
Me
297.8
27.7
115.0
9.7
102
232
198
11
13
1.8
3.9
1.6
31
35
CF
CF
3
Me
>300
3
a) PDE2A IC data is from resynthesized samples, not from initial library compounds.
5
0
A number of analogs wherein the 5ꢀaryl pyrazole group had differing substitution at the paraꢀ
position were prepared in the library which showed larger substituents provided increased
potency (Table 9). These compounds were resynthesized and two additional compounds (35 and
3
6) were prepared. The unsubstituted phenyl (36) had PDE2A IC = 566.1 nM. Progressively
50
larger substituent size led to significant increases in potency. For example, addition of a fluoro
37) gave a 2ꢀfold potency increase, methyl (38) yielded a 20ꢀfold potency increase, while the
(
larger trifluoromethyl (35) and ethyl (40) resulted in 57ꢀfold and 80ꢀfold potency increases,
respectively. Both LE and LipE increased with increasing substituent size, suggesting that the
potency was being driven by more than nonꢀspecific lipophilic interactions which would result in
lower LE and LipE values.
Table 9. Impact of 5ꢀaryl paraꢀposition substitution and impact on PDE2A potency and binding
efficiency measures
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
PDE2A
Cpd. ID
R
a
clogP
LE
LipE
IC (nM)
5
0
36
37
38
H
F
MeOꢀ
Me
CF
Et
566.1
240.8
108.5
27.7
9.7
2.8
3.0
2.8
3.3
3.7
3.8
0.34
0.35
0.35
0.4
0.38
0.41
3.4
3.6
4.2
4.2
4.3
4.3
34
35
3
39
7.1
a) PDE2A IC data is from resynthesized samples, not from initial library compounds.
5
0
A protein xꢀray crystal structure of 39 bound to PDE2A was obtained to help clarify the roles
that the ether, methylated pyrazole, and phenyl paraꢀsubstitution all play in potency enhancement
(
Figure 4a). First, the triazolopyrazine core of 39 maintains the same interactions with the
enzyme as 3. The nꢀpropyl ether occupies the same space as the dimethoxyphenyl group in 1,
while avoiding deleterious interactions with Leu770. The 1ꢀmethylꢀ5ꢀarylpyrazole in 39
overlaps with the 2ꢀhydroxyꢀ6ꢀphenylhexanꢀ3ꢀyl sidechain in 1. The sp2 nitrogen (Nꢀ2) of the
pyrazole is wellꢀaligned with the hydroxyl in 1, accepting a hydrogen bond from a water
molecule. The vector to access the induced pocket off the pyrazole 5ꢀposition with a rigid phenyl
linker allows para substituents to reach the deep recesses of the pocket, with a lower entropic
penalty than in the conformationally flexible sidechain of 1.
a)
b)
ACS Paragon Plus Environment

Page 21 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. a) Xꢀray crystal structure of 39 bound to PDE2A. b) Overlap of 1 and 39 in PDE2A.
In terms of understanding the increase in potency observed when methylating the pyrazole, the
methyl does not appear to make any specific interactions in the pocket nor does it appear to
displace any water molecules. Dihedral angle scans were performed on three truncated phenyl
pyrazole systems (40, 41, and 42, Figure 5), which suggest that the methyl on the pyrazole
nitrogen has an effect on the orientation of the phenyl ring. The Xꢀray coꢀcrystal structure of 39
with PDE2A reveals a preferred average dihedral angle of 67° in the bound conformation over
the four structures in the asymmetric unit. The local minimum energy structures of 40 and the
pyrazole tautomer 42 have dihedral angles of 35.3° and 28.6°, respectively, while that of 41 has a
dihedral angle of 45.7° (see Supporting Information). This suggests that the methyl on the
pyrazole ring causes the phenyl ring to adopt a conformation which is closer to that observed in
the Xꢀray coꢀcrystal structure, leading to a smaller conformational energy penalty upon binding.
Optimization of the full systems and the dihedral scans of the truncated systems indicate that the
conformational strain energy required to achieve the bound conformation for compounds without
a methyl on the pyrazole, ~1.4 – 2.4 kcal/mol (depending on the tautomeric mixture), is at least
twice that of those with a methyl on the pyrazole, ~0.6 kcal/mol.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Truncated systems used to calculate dihedral angles in ground state.
Compound 39 was profiled more extensively for ADME properties and PDE selectivity (Table
10). The selectivity of this compound against other PDEs was exceptionally high, driven most
likely by efficient filling of the PDE2A induced pocket. HLM Cl was high, but low MDR1
int
BA/AB efflux was observed. While the PDE2A potency was improved relative to 3, the high
HLM Clint negatively impacted estimated human dose predictions affording a value of >1438
mg/day.
Table 10. Comparison of PDE2A potency, selectivity and ADME profiles of 39 and 3
Cpd. ID
39
3
PDE2A IC (nM)
7.1
390
5
0
<
20x
>2400x
(1B/3A/7B/10A/11A);
9X (4D); 71x (5A); 69x
PDE Selectivity
(1B/3B/4D/5A/6/7B/8B/
2
9A/10A/11A)
(6); >200x (8B/9A)
HLM Clint*
MDR1 BA/AB
>300
1.85
>287
1.2
a
estimated human dose (mg/day)
>1438
>46136
a) estimated human dose (mg/day) = 10 x PDE2A IC x HLM CL x 24 h x 70 kg x MDR1 BA/AB.
50
int
ACS Paragon Plus Environment

Page 23 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In order to address the high estimated human dose, improving PDE2A potency and reducing
HLM Cl were targeted. Overlaying protein xꢀray crystal structures of 4 and 39 showed that the
int
heterocyclic cores aligned with one another regarding the substituents that projected into the
ribose region, which were the isobutyl of 4 and aryl pyrazole of 39 (Figure 6a). A hybrid series
was proposed which merged the pyrazolo[3,4ꢀd]pyrimidine core with the aryl pyrazole side
chain in 39 (Figure 6b).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(a)
MeO
O
N
N
N
O
+
N
N
N
N
N
N
N
N
39
4
N
N
R'
R
N
N
N
N
N
43
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
b)
Figure 6. a) Overlap of 4 (yellow) and 39 (grey) in PDE2A. b) Merging 4 and 39 to yield a
proposed hybrid series 43.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In order to access compounds of general structure 43, pyrazole boronate 44 was coupled with
an aryl bromide (45) to afford 46 (Scheme 8). Formylation of 46 with dimethylformamide and
phosphorous oxychloride yielded the 4ꢀformyl pyrazole 47. Low temperature lithiation of 4,6ꢀ
dichloropyrimidine (48) at the 5ꢀposition followed by addition of 47 gave the carbinol 49.
Oxidation of 49 yielded the corresponding ketone 50 which could be reacted with an amine
(
HNRR’) at room temperature to selectively provide the monoꢀamino product 51. Subsequent
reaction with methyl hydrazine afforded the target compounds of general structure 44.
Br Ar
45
b
N N
N N
N N
a
Ar
B(OH)2
Ar
(5)
(
6)
H
O
44
46
47
N N
N N
Ar
Cl
N
Ar
Cl
N
Cl
c
4)
d
N
N
O
N
OH
(
(3)
Cl
Cl
N
Cl
48
49
50
e
(
2)
N
N N
N
NRR'
f
Ar
NRR
Ar
N
(
1) N
O
N
N
N
N
Cl
51
44
ACS Paragon Plus Environment

Page 25 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 8. a) PdCl (PPh ) , NaHCO , DME, H O; b) DMF, POCl ; c) i. 48, LDA, THF, ii. 47 d)
2
3 2
3
2
3
DessꢀMartin periodinane; e) HNRR’, MeCN; f) MeNHNH , pyridine.
2
Initial analogs with varied amines were prepared via a 2ꢀstep parallel chemistry protocol (50
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ
51 ꢀ 44, Scheme 8). Trends for potency were demonstrated in this initial library with
azetidine and 3ꢀsubstituted azetidines providing the most potent PDE2A inhibitory activity
Table 11, 52-54), comparable to or better than 39. With lower clogP relative to 39 (ca. 1 unit),
(
HLM Cl was reduced for these compound while maintaining low MDR1 BA/AB efflux which
int
reduced the estimated human dose for 52 and 53 (160 and 337 mg/day, respectively). The polar
methyl carbamate of 54 also showed reduced HLM Cl . However, with significant MDR1
int
BA/AB efflux, the estimated human dose increased (954 mg/day). Overall, LE and LipE values
were also in a desirable range, especially for 52 (LE = 0.43, LipE = 6). The reduced
conformational flexibility of 52 (RB = 4) most likely resulted in a low energy ground state that is
similar to the PDE2A enzyme bound conformation, obviating an entropic penalty for binding,
which is associated with the more flexible 1. Pyrrolidines (55-56) provided slightly less activity
than azetidines, but had higher human dose due to greater MDR1 BA/AB efflux. Sixꢀmembered
ring amines (57-58) showed significantly less PDE2A activity.
Table 11. Representative parallel synthesis compounds varying 5ꢀamino substituent
Estimated
Cpd.
ID
PDE2A IC
HLM MDR1
Clint BA/AB
human
dose
(mg/day)
5
0
ꢀNRR
a
clogP
LE
LipE
(nM)
b
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
5
2
3
2.0
5.8
150
77
1.4
1.9
160
337
2.7
2.9
0.43
0.38
6
5.4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
4
3.6
42
14
954
2.8
3.3
0.36
0.39
5.6
4.8
c
55
7.8
200
3.1
1859
5
6
4.9
55
5.4
642
2.9
0.36
5.5
57
423.0
149
60
2.2
6.2
55,657
2.5
2.9
0.30
0.26
3.9
3.0
5
8
1407.0
222,323
a) PDE2A IC50 data is from library compounds. b) estimated human dose (mg/day) = 10 x PDE2A IC50
x
HLM, Clint x 1440 min x 70 kg x (MDR1 BA/AB). c) PDE2A IC50 is the geometric mean of two
independent IC50 measurements.
An xꢀray crystal structure of 56 bound to PDE2A was successfully solved (Figure 7).
Consistent with the PDE2A xꢀray crystal structure of 4, the pyrazolopyrimidine core made the
same polar and nonꢀpolar interactions with the protein and the 1ꢀmethylꢀ5ꢀarylpyrazole moiety
resided in a similar position as in 39. While azetidine and pyrrolidine were accommodated at the
4
ꢀposition of the pyrazolopyrimidine, it is possible that the bulkier sixꢀmembered ring systems
ACS Paragon Plus Environment

Page 27 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
such as morpholine and piperazine cannot fit as well and make unfavorable contacts with sideꢀ
chains lining the bottom and sides of the binding pocket (Ileꢀ826, Pheꢀ830, Metꢀ847).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Xꢀray crystal structure of 56 bound to PDE2A
52 made via parallel synthesis was resynthesized on larger scale and more fully profiled. The
PDE2A IC for this resynthesized material was equivalent to the compound made via parallel
5
0
synthesis. In an analogous manner to 39, selectivity against other PDE families was >3000x,
providing further support that efficient binding to the induced pocket in PDE2A could be a
general approach to minimizing offꢀtarget PDE activity.
The focus was nexted placed on exploring changes to the pyrazole 5ꢀaryl group to reduce HLM
Cl while maintaining low MDR1 BA/AB. The route in Scheme 8 was used to synthesize other
int
substituted phenyl derivatives. An alternative route was developed (Scheme 9) to allow the
incorporation of heterocycles which were lowꢀyielding in the original method. Selective S Ar
N
of azetidine on 59 proceeded in high yield to afford 60. The heteroaryl pyrazole moiety was
constructed via Suzuki coupling between pyrazole boronic acid 61 and the haloꢀheterocycle 62 to
give 63. Direct iodination of 63 proceeded cleanly and in high yield to give 64, which could be
metallated with either nꢀBuLi or iꢀPrMgCl and then treated with 60 to yield the carbinol 65.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Oxidation then gave the 4,6ꢀdichloroꢀ5ꢀketoꢀpyrimidine 66 (similar to 51) which readily reacted
with methyl hydrazine to provide the target compounds 67.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 9. Synthetic route to heteroaryl substituted pyrazole analogs 67. a) azetidine, DIPEA,
CHCl , 64%; b) PdCl (PPh ) , Na CO , 1,2ꢀDME, H O; c) NIS, AcOH; d) nꢀButyl lithium or i-
3
2
3 2
2
3
2
PrMgCl, THF; e) DessꢀMartin periodinane, CH Cl ; f) NH NHMe, pyridine.
2
2
2
Potency trends with regard to aryl paraꢀsubstitution tracked with the triazolopyrazine results
(compare Table 9 to Table 12). For example, adding a 4ꢀmethyl to the parent phenyl compound
(68 ꢀ 69) gave an order of magnitude enhancement in potency and higher LE, both of which
were further increased in the case of ethyl (52). Similarly, cyclopropyl (70) and trifluoromethyl
(71, PFꢀ05085727) also provided high potency. These more potent compounds maintained low
MDR1 BA/AB with HLM Cl still in the moderate to high range, with 71 having an estimated
int
human dose <100 mg/day. The replacement of phenyl with pyridine did reduce HLM Cl 2ꢀ3
int
ACS Paragon Plus Environment

Page 29 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
fold when looking at matched pairs (71 vs. 72; 70 vs. 73), but potency was reduced significantly
(3ꢀ13 fold) which ultimately led to increased estimated human dose. Shifting from pyridine to
the more polar pyrimidine (73 ꢀ 74) reduced HLM Cl 2ꢀfold but afforded a 100ꢀfold loss in
int
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
potency along with increased MDR1 BA/AB efflux.
Table 12. SAR of pyrazole 5ꢀaryl group
estimated
Cpd.
ID
PDE2A IC50 HLM MDR1
human
ꢀR
a
clogP
LE
LipE
(nM)
Clint BA/AB
dose
(mg/day)
6
8
188.7
34
47
1.1
1.8
1.4
1.4
1.4
2512
1.7
2.2
2.7
2.6
2.6
0.35
0.40
0.43
0.41
0.41
4.6
5.5
6
b
69
17.7
546
160
166
72
5
2
2.0
3.4
1.6
150
90
70
5.9
6.1
7
1
77
72
73
74
8.9
8.9
33
53
28
2.3
3.1
5.6
282
565
1.5
1.4
0.37
0.39
0.29
6.5
6.6
5.5
1012.7
60759
0.48
a) Estimated human dose (mg/day) = 10 x PDE2A IC x HLM, CL x 1440 min x 70 kg x (MDR1 BA/AB).
50
int
b) PDE2A IC is the geometric mean of two independent IC measurements.
50
50
In order to explore in vivo PDE2A inhibition, 71 was chosen to be profiled extensively to assess
its potential for use as an in vivo tool. Key to choosing this compound was its lower rat liver
ACS Paragon Plus Environment